Improving Uptake of Cancer Genetic Risk Assessment in a Remote Tailored Risk Communication and Navigation Intervention: Large Effect Size but Room to Grow

PURPOSE

Cancer genetic risk assessment (CGRA) is recommended for women with ovarian cancer or high-risk breast cancer, yet fewer than 30% receive recommended genetic services, with the lowest rates among underserved populations. We hypothesized that compared with usual care (UC) and mailed targeted print (TP) education, CGRA uptake would be highest among women receiving a phone-based tailored risk counseling and navigation intervention (TCN).

METHODS

In this three-arm randomized trial, women with ovarian or high-risk breast cancer were recruited from statewide cancer registries in Colorado, New Jersey, and New Mexico. Participants assigned to TP received a mailed educational brochure. Participants assigned to TCN received the mailed educational brochure, an initial phone-based psychoeducational session with a health coach, a follow-up letter, and a follow-up navigation phone call.

RESULTS

Participants' average age was 61 years, 25.4% identified as Hispanic, 5.9% identified as non-Hispanic Black, and 17.5% lived in rural areas. At 6 months, more women in TCN received CGRA (18.7%) than those in TP (3%; odds ratio, 7.4; 95% CI, 3.0 to 18.3; P < .0001) or UC (2.5%; odds ratio, 8.9; 95% CI, 3.4 to 23.5; P < .0001). There were no significant differences in CGRA uptake between TP and UC. Commonly cited barriers to genetic counseling were lack of provider referral (33.7%) and cost (26.5%), whereas anticipated difficulty coping with test results (14.0%) and cost (41.2%) were barriers for genetic testing.

CONCLUSION

TCN increased CGRA uptake in a group of geographically and ethnically diverse high-risk breast and ovarian cancer survivors. Remote personalized interventions that incorporate evidence-based health communication and behavior change strategies may increase CGRA among women recruited from statewide cancer registries.

BACKGROUND

Identifying inherited cancer can help guide clinical care and improve outcomes through prevention, early detection, and targeted treatment.^1-5 Up to 20% of cancers have an inherited basis, including those of the breast and ovary.^6,7 Pathogenic variants in the BRCA1 and BRCA2 genes account for most hereditary breast and ovarian cancers (HBOCs), followed by genes such as CHEK2, PALB2, and others.^1,7-10 Pathogenic variants in these genes, particularly BRCA1/2, also increase the risk for cancers in family members (eg, breast, ovarian, pancreatic, and prostate cancers). National guidelines recommend cancer genetic risk assessment (CGRA) for all women with epithelial ovarian, fallopian tube, primary peritoneal, and high-risk breast cancer.¹¹ CGRA involves a clinical assessment of a person's risk and counseling and/or testing where appropriate. Although the clinical utility of CGRA is well established, many patients who meet national criteria have not had genetic counseling and/or testing,¹¹ with the lowest rates among minority, rural, and other underserved populations.^11-15 Predictors of suboptimal CGRA utilization include inadequate risk identification, lack of awareness, psychosocial factors, language barriers, cost concerns (including lack of insurance), and lack of provider referral.^15-18

CONTEXT

• Key Objective

• Although the clinical utility of cancer genetic risk assessment (CGRA) is well established, many patients who meet national criteria have not had genetic counseling and/or testing, with the lowest rates among minority, rural, and other underserved populations.

• Knowledge Generated

• In this three-arm trial, women in the phone-based tailored counseling and navigation (TCN) intervention group were significantly more likely than those in the targeted print (TP) and usual care groups to obtain CGRA (ie, genetic counseling and/or testing) by the 6-month follow-up assessment. TCN was effective in Hispanic and rural women and women with low health literacy.

• Relevance (S.B. Wheeler)

• Phone-based tailored counseling with navigation is effective in improving rates of CGRA, notably in groups that historically are underserved, including Hispanic, rural, and low-literacy populations.*

• *Relevance section written by JCO Associate Editor Stephanie B. Wheeler, MPH, PhD.

Many cancer survivors remain unaware of their HBOC risk.¹⁹ This represents a missed opportunity to inform women and their at-risk relatives about their future cancer risks. Addressing this translational gap is crucial to evidence-based cancer prevention and risk management strategies that seek to achieve a population-level reduction in cancer morbidity, mortality, and disparities.^20,21 This lost opportunity to identify families with pathogenic variants in cancer risk genes and prevent new cancers led the National Cancer Institute to call for studies evaluating the impact of a novel, but untested, genetic testing strategy called Traceback.^19,22 Traceback aims to find high-risk families by identifying pathogenic variant carriers who did not receive testing at the time of their diagnosis and are thus unaware of their elevated risk.²²

The Genetic Risk Assessment for Cancer Education and Empowerment (GRACE) Project addressed this need by leveraging statewide cancer registries as a low-cost approach^23-25 to deliver a remote, telehealth intervention for ethnically and geographically diverse high-risk cancer survivors. The purpose of the GRACE Project was to inform participants of their potentially increased risk and motivate and navigate them to receive a CGRA. The GRACE Project's phone-based tailored counseling and navigation intervention incorporated health communication and behavior change theories alongside risk messages and motivational strategies to address key risk factors and increase uptake of CGRA. Similar approaches have been effective in decision making for cancer screening and treatment, but to our knowledge, this is the first population-level study evaluating the impact of decision coaching and navigation on CGRA uptake. Informational print materials (eg, brochures) have been widely used as a high-value public health strategy to promote cancer prevention behaviors. Although access to digital health information has been increasing, many patients still prefer and use print information, especially persons from underserved populations.^26-28 Cancer survivors from diverse socioeconomic, racial, and ethnic groups cite print educational communications as a preferred method of delivering information, including information about CGRA.^17,26,29-32 However, personalized psychoeducational interventions that inform, motivate, and address barriers to clinical services are typically more effective than nonpersonalized interventions in facilitating access to recommended health services.³³ Evidence suggests that activated and engaged patients equipped with the necessary skills are more likely to follow through with recommended care and have better health outcomes at reduced costs.^34-37

METHODS

Study Design

GRACE was a randomized three-arm superiority trial testing the impact of three conditions on CGRA uptake among ovarian and high-risk breast cancer survivors: (1) a tailored (personalized) remote intervention that combined evidence-based decision coaching and behavior change techniques with navigation (TCN); (2) a nontailored (targeted) print brochure promoting CGRA (TP); and (3) usual care (UC).³⁸ We hypothesized that CGRA uptake at 6 months (as verified by medical record documentation of genetic counseling and/or testing) would be higher among participants who received TCN, compared with participants who received TP or UC. We report this trial following the recommended standards of the extended Consolidated Standards of Reporting Trials statement for parallel-group, nonpharmacologic randomized trials.^39,40 GRACE was approved by the participating institutions' Institutional Review Boards, and all participants provided informed consent. Recruitment began in November 2017 and ended in July 2020. Primary outcome assessments (CGRA 6-month follow-up) were completed in February 2021.

Participants

To meet recruitment goals, we invited all potentially eligible women from the sampling frames of three statewide cancer registries in Colorado, New Jersey, and New Mexico. Consistent with National Comprehensive Cancer Network guidelines, eligibility included female sex, diagnosed with breast cancer at age ≤ 50 years, triple-negative breast cancer diagnosed at age ≤ 60 years, and/or two or more primary breast cancers. Women with epithelial ovarian, peritoneal, and/or fallopian tube cancer were also eligible. Women were eligible if they were age ≥ 21 years, resided in Colorado, New Mexico, or New Jersey at the time of enrollment, and could read and speak English and/or Spanish fluently. Women were excluded if they reported prior genetic counseling or testing, were unable to give informed consent, or did not have access to a telephone.

Random Assignment and Masking

After confirmation of eligibility, informed consent, and completion of baseline surveys, a project coordinator randomly assigned (1:1:1) participants to one of the three study arms, using a computer-generated allocation algorithm on the basis of block random assignment with nine patients per block, which provided an approximate balance between study arms and within racial and ethnic and geographic strata.

Data Collection, Measures, and Primary Outcome Assessment

Cancer diagnosis, including pathology, was obtained from the cancer registries while other eligibility criteria were collected by self-report over the telephone. Evidence-based procedures were used to collect baseline and outcome data, including $50 incentives after completion of follow-up surveys.⁴¹ Data collectors were blinded to intervention assignment. Participants completed surveys online or by telephone (on the basis of preference) at baseline, 1 month, and 6 months after the interventions and for the UC arm at baseline, 1 month, and 6 months after random assignment. Participants who reported receiving a CGRA (ie, genetic counseling and/or testing) were asked to provide written consent to obtain documentation of receipt of these health services. The primary outcome was medical record–verified CGRA. Psychological factors were assessed with the following three validated scales.

• Fear of HBOC was measured with the six-item Cancer Risk Beliefs Scale's Negative Affect in Risk subscale that captured participants' fear and feelings about their risk for HBOC⁴² (Cronbach's α = 0.94).

• Cancer worry⁴³ was measured with a three-item scale that measured the intensity and frequency of cancer worry (α = 0.91).

• Health literacy⁴⁴ was assessed with a validated three-item measure that is used to screen for people with inadequate health literacy (α = 0.74).

Study Arms

Participants randomly assigned to UC received the surveys; no intervention was delivered. The intervention conditions are briefly described below, in Figure 1, and in more detail elsewhere.³⁸

FIG 1.

Graphical depiction of TP and TCN interventions. (a) Baseline survey, which captured sociodemographic information (age, rural status, etc), self-reported health status, cancer diagnosis, health literacy, number of living first degree relatives, cancer worry, perceptions of threat and efficacy, presence of a primary care provider and/or cancer care provider, history of provider CGRA recommendation, genetic counseling, genetic testing, barriers, and facilitators; (b) notification letter indicating random assignment and next steps; (c) mailed educational brochure; (d) sealed envelope of visual aids sent by mail; (e) TCN telephone session with a cancer education specialist addressing perceived hereditary breast and ovarian cancer risk, threat, response efficacy, and self-efficacy; creation of action plan to obtain CGRA; and navigation assistance to overcome specific barriers to CGRA; (f) mailed tailored summary letter of TCN telephone session and outlined the participant’s stated initial steps to getting CGRA. With the participants permission, a copy was mailed to the patient’s provider; (g) mailed tailored reminder card detailing genetic counseling and genetic testing action plan; and (h) follow-up call from cancer education specialist (for those who verbally consented to a call at the end of the initial TCN session). CGRA, cancer genetic risk assessment; HBOC, hereditary breast and ovarian cancer; TCN, tailored counseling and navigation; TP, targeted print.

Our goal was to develop print materials at < 6th-grade readability level. However, some essential terms inflated the readability of some sections resulting in an 8th-grade readability overall.

Targeted print.

Participants randomly assigned to TP received a mailed, study-specific educational brochure that was targeted to their genetic risk status and addressed evidence-based theoretical targets: awareness of CGRA guidelines (knowledge), threat appraisal (validate or raise risk perceptions of HBOC seriousness), response efficacy (benefits and expectations about CGRA), and self-efficacy (CGRA resources, locations, and insurance reimbursement). The TP brochure provided state-specific information about where to access clinical cancer genetic services.

Tailored counseling and navigation.

Participants randomly assigned to TCN received the mailed educational brochure, followed by a 30-45 minute phone call with a health coach. Coaches were trained in motivational interviewing (MI), an evidence-based communication style,⁴⁵ and received supervision by an experienced MI trainer. The session incorporated risk communication and behavior change techniques to raise perceptions of threat of HBOC, and on the basis of implementation-intention tenets,^46,47 participants were asked to create an action plan for obtaining genetic counseling. The basic format of the session involved (1) providing information about the structure of the session, the role of the health coach, and establishing rapport; (2) elicitation of behaviors the patient was already using to address their cancer risk; (3) tailored information about the participant's risk; (4) information about the purpose and benefits of CGRA; (5) assessment of the participant's motivation and readiness to obtain a CGRA; and (6) assistance with action planning and navigation to a cancer genetic clinic where appropriate. Participants were reminded of their action plan in a follow-up tailored letter (sent immediately after the intervention) that summarized key intervention targets, the action plan, and navigation strategies, as needed. With the participant's permission, the health care provider received a letter informing them that the participant met the referral criteria for CGRA and a copy of the tailored letter. Finally, participants received a 10-20 minute follow-up phone call approximately 7 weeks after the initial call to assess the need for additional navigation and assistance.

Data Analysis

The primary outcome analyses were performed on an intent-to-treat basis. Between-group differences in demographic and clinical variables were assessed using ANOVA and χ² tests. Logistic regression analysis, with CGRA uptake (yes/no) as the outcome variable, was employed to compare the treatment effect between TCN versus UC and TP as the primary analysis. Comparison between UC and TP was also performed as a secondary analysis. We also tested the effect of the interventions for those with a self-reported outcome and used both negative and multiple imputation methods to estimate CGRA uptake in those with an unknown outcome as sensitivity analyses to determine whether the intervention effect estimates were sensitive to imputation and if the conclusions were consistent. Specifically, negative outcome imputation assumed that if there was no documented verification of CGRA, the procedure did not occur. Multiple imputation^48-50 assumed missing at random and was based on verified CGRA within 6 months and demographic information, including age, race, and ethnicity; marital status; health insurance status; education level; health literacy level; employment status; household income; rural versus urban residence; having a primary care provider; cancer site (breast v ovarian cancer); and number of first- and second-degree relatives with cancer, as well as the a priori interaction of the intervention with potential effect modifiers: race, ethnicity, health insurance status, literacy level, and household income. Ten imputed data sets were generated using IVEware⁵¹ which uses the sequential regression multivariate imputation approach to multiply imputing item missing values. Rubin's rule⁵⁰ was subsequently applied to combine results using SAS Proc Mianalyze. For all analyses, we calculated odds ratio (OR) estimates, percentages of CGRA, and the 95% CIs. Analyses were conducted using SAS v9.4 (SAS Institute Inc, Cary, NC).

On the basis of the final sample size of at least 212 per arm and about 3% of CGRA for TP or UC, the minimum detectable OR for comparisons between TCN and TP or between TCN and UC is 1.86 with 80% power and 5% overall type I error rate (2.5% for each comparison) after multiplicity adjustment. Refer statistical analysis plan in the Data Supplement (online only) for more details.

RESULTS

The study enrollment, random assignment, and retention data are shown in Figure 2. Of those deemed to be eligible at the initial assessment (n = 821), 668 participants were randomly assigned to a study arm; 27 were found to be ineligible after random assignment because they were later found to have had a CGRA. Eligible women who underwent random assignment (n = 641) were included in the analysis. Seventy-four percent of TCN sessions were evaluated for treatment integrity with 95% fully compliant with the intervention fidelity checklist. Key indicators of MI fidelity (eg, MI global scores, % complex reflections, and reflection to question ratio) were consistently in the good to excellent range. The retention rate was 91.3%, with no statistically significant differences in the percentage of participants from each arm who were available at 6 months for assessment of CGRA uptake. Participants whose 6-month outcome data on key demographic and clinical factors were unknown did not differ from those whose outcome data were known. Participants in the three arms did not differ in baseline demographic or clinical characteristics (Table 1).

FIG 2.

GRACE study CONSORT diagram. GC/GT, genetic counseling and/or testing; GRACE, Genetic Risk Assessment for Cancer Education and Empowerment; MVR, medical record verified; TCN, tailored counseling and navigation; TP, targeted print. ^aIn the intent-to-treat analysis, only participants found ineligible after random assignment were excluded.

TABLE 1.

Sociodemographic and Clinical Characteristics of Participants by Study Arm

Of the 43 women who had CGRA, 65% had both counseling and testing, 24% had testing only, and 12% had counseling only. The percentages of patients who obtained CGRA by the 6-month follow-up are presented in Table 2. A higher percentage of CGRA in TCN was consistently observed, regardless of the imputation method. A logistic regression analysis (Table 3) showed that higher CGRA rates occurred among participants randomly assigned to the TCN study arm compared with the TP arm (OR, 7.4; 95% CI, 3.0 to 18.3) and UC arm (OR, 8.9; 95% CI, 3.4 to 23.5). CGRA rates were not significantly different between TP and UC. Comparisons of TCN's effect were robust and consistent, regardless of the imputation method. The most frequently cited barriers to counseling were lack of provider referral (33.7%) and cost concerns (26.5%) and regarding genetic testing were cost concerns (41.2%) and anticipated challenges with coping with results (14.0%).

TABLE 2.

Self-Reported and Medically Verified Cancer Genetic Risk Assessments by Study Arm

TABLE 3.

Logistic Regression Model Results for Intervention Effects on CGRA Within 6 Months

DISCUSSION

Compared with UC and mailed TP, TCN increased CGRA uptake at 6 months. To our knowledge, this is the first population-based randomized controlled trial to use a Traceback approach that incorporates remote, personalized psychoeducation and navigation to increase CGRA uptake among a group of ethnically and geographically diverse high-risk cancer survivors. A prior study that used an educational brochure and video to increase CGRA uptake among breast cancer survivors improved motivation but did not appreciably improve CGRA uptake.⁵² Tailored approaches have been shown to be more effective than generic educational materials in promoting behavior change,^33,34,53-56 and patient navigation has been shown to reduce disparities and increase access to care.^57,58 In part, the impact of TCN may stem from its theoretical grounding in risk communication and MI to increase motivation and commitment to obtain CGRA. MI itself has a substantial track record of helping people change behaviors, including breast and cervical cancer screening. Tailored action planning and navigation likely also bolstered the intervention^47,59; however, our study was not designed to disentangle the unique effects of each intervention component. Although TP was also designed to address key theoretical determinants of CGRA uptake (eg, informational needs, threat/risk and efficacy), the absence of an effect may be due to the lack of personalization in the print materials and the absence of a health coach who can facilitate a customized discussion of risks, motivation, barriers, and plans for obtaining CGRA.

Traceback case ascertainment is a critical component of reaching diverse populations because of known disparities in referral patterns to CGRA and linguistic barriers. This underscores the importance of TCN's findings, particularly compared with other studies that have attempted to increase CGRA and other types of cancer screenings in samples of primarily urban dwellers and non-Hispanic White patients.^52,60

Low CGRA uptake is common in clinical, research, and direct-to-consumer testing contexts. It is estimated that up to 87% of women who are referred for genetic counseling by their physician do not follow through with the referral.⁶¹ Furthermore, only 0.2% of 4,788 patients in a genomic sequencing clinical study took advantage of elective genetic counseling before testing.⁶² Interventions like ours that are designed to improve people's knowledge of cancer risk, testing options, and navigation assistance may be one way to improve rates of genetic counseling and testing. However, at the same time, we note that only 19% of the women in TCN received a CGRA. Because of the geographic spread of our sample (and their recruitment from state cancer registries and affiliation with myriad health systems) and centralized health coaches, we were not able to directly schedule people for CGRA or estimate/arrange payment for services (two commonly mentioned barriers). Better integration of a relatively low-intensity, low-cost intervention such as TCN with other clinical services represents an opportunity for improvement in future studies. Additionally, the standard approach of referral to pretest genetic counseling^63,64 has been cited as a common barrier to receiving guideline-concordant genetic testing,^65,66 especially for underserved minorities and low socioeconomic status survivors.⁶⁷ Our relatively low uptake rates may highlight the need to re-evaluate the requirement in many health systems for patients to pursue a pregenetic test consultation with a certified genetics professional. The goal to expand access highlights the need to develop and test alternative strategies to educate, support, and connect cancer survivors to genetic education, counseling, and testing.^68-70

Our study included a diverse sample of participants from underserved populations, relative to a similar study, which primarily enrolled non-Hispanic White population with at least some college education.⁵² Much of this is owed to our Traceback^19,22 approach using cancer registries for recruitment. All activities were conducted in English and Spanish and drew from New Mexico, Colorado, and New Jersey, states that are ethnically, geographically, and linguistically diverse. Consequently, Hispanic and rural women were included, groups that historically have had limited access to CGRA.^71,72 Based on state demographics, we believe we would have been able to recruit even more Hispanic women; however, concerns about documentation status likely represented a barrier to enrollment for some women who qualify for CGRA.^73,74

Limitations of our study include the relatively small number of Black women who enrolled. Because of our population-based approach, we were not able to carefully evaluate barriers to enrollment. We urge future research to evaluate strategies to expand access to cancer genetic services among Black women and other underserved and understudied populations. Our study was not powered to conduct a meaningful subgroup analysis to determine if TCN and TP had similar effects across racial, ethnic, and geographic subgroups.

Overall, our findings support the expanded use of personalized remote risk communication interventions such as TCN to increase CGRA uptake among women at increased risk of HBOC. We believe TCN's theoretically guided approach to patient education and empowerment combined with navigation creates a promising approach to enhance motivation and remove barriers. TCN represents an important step in reducing disparities in genetic care without overburdening scarce genetic counseling resources.

DATA SHARING STATEMENT

The following data will be shared with researchers who provide a methodologically sound proposal for meta-analyses beginning 12 months years following article publication (no end date): individual participant data that underlie the results reported in this article after deidentification, informed consent document and study Protocol (online only). Proposals should be directed to anita.kinney@rutgers.edu. To gain access, data requesters will need to sign a data access agreement.

AUTHOR CONTRIBUTIONS

Conception and design: Anita Y. Kinney, Scott T. Walters, Lisa E. Paddock, Jean A. McDougall

Provision of study materials or patients: Antoinette Stroup, Lisa E. Paddock, Sherry Grumet

Collection and assembly of data: Anita Y. Kinney, Scott T. Walters, Arreum Kim, Julianne Ani, Emily Heidt, Antoinette Stroup, Lisa E. Paddock, Sherry Grumet, Tawny W. Boyce, Jean A. McDougall

Data analysis and interpretation: Anita Y. Kinney, Scott T. Walters, Yong Lin, Shou-En Lu, Emily Heidt, Circe J.G. Le Compte, Lisa E. Paddock, Deborah L. Toppmeyer, Jean A. McDougall

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Improving Uptake of Cancer Genetic Risk Assessment in a Remote Tailored Risk Communication and Navigation Intervention: Large Effect Size but Room to Grow

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).